FEATURED
06:20 UTC
Cursor 3.0 pushes an agent SDK; devs flag gaps as Windsurf gains ground
trend pattern
medium
Agent IDEs are maturing, but standardize only after a measured bake-off; the SDK and guardrails will decide your winner.
claude-code
06:20 UTC
Claude Code 2.1.136–137 lands enterprise guardrails and stability fixes (Windows, MCP, OTel)
new feature deep dive
medium
Update to 2.1.137 for Windows stability and use 2.1.136’s hard_deny and OTel survey support to tighten governance and observability.
anthropic
06:21 UTC
Anthropic lifts Claude Code caps and throttles, backed by SpaceX Colossus capacity
new feature deep dive
medium
Claude Code finally behaves like a dependable backend service: higher limits, fewer throttles, and smoother long sessions.
swe-bench-verified
06:22 UTC
Context beats model: a cheap agent tops SWE-bench Verified
data benchmark study
medium
Better context pipelines can beat bigger models—optimize the agent harness before you upgrade the LLM.
openai
06:23 UTC
Codex goes headless (remote-control server) and into Chrome; reports of idle credit drain surface
release problems outages controversies
medium
Codex is ready for centralized, browser-accessible agent workflows—test cost behavior before scaling.
github
06:25 UTC
Copilot CLI 1.0.44 makes terminal agents more programmable; Copilot billing shift looms
new feature deep dive
medium
Copilot CLI just got programmable short-circuits for faster, cheaper automation; usage-based Copilot billing is next—align your workflows and budgets now.
anthropic
06:26 UTC
Anthropic’s Project Glasswing puts AI vuln discovery into production (with a path to auditability)
new product launch
high
AI-first vulnerability discovery is arriving in production; pilot it now and pair it with strict data boundaries and audit trails.
anthropic
06:27 UTC
Machines just out-reviewed us: 271 Firefox bugs and the AI security pivot
trend pattern
high
Plan for AI-first security: integrate machine audits into CI and threat-model agent tool/memory surfaces before they bite.
postgresql
06:30 UTC
claude-mem goes server-side (and Apache-2.0): open-source agent memory grows up
new feature deep dive
medium
Agent memory just moved from plugin to platform — and the license now says “yes” to enterprise pilots.
claude
06:31 UTC
Smarter Claude agents are burning 54% more tokens; the fix is backend context, not a bigger model
data benchmark study
medium
Better models won’t fix missing context—give agents a small, complete backend manifest and scoped skills to cut token burn and retries.
chrome
06:35 UTC
Codeex Browser Agent adds natural-language control of Chrome
new feature deep dive
medium
Prompt-driven, self-healing Chrome automation is now on the table—try it on your flakiest web workflow.